English summary: Prescribing for Children: From Art to Science
Introduction
It is generally acknowledged that a substantial proportion of prescriptions for
children are written in the absence of sound scientific evidence regarding the
effectiveness of the drugs. Recent publications have demonstrated that for many
drugs, the majority of prescriptions for children are unlicensed or off-label.
Clinical research on children in order to obtain this scientific evidence would be
expensive, pose ethical issues, and be complicated (involving assessments of
dosage level, side effects, and considerations of different age levels). Because
research to obtain such evidence for use in the paediatric population is not
required for a new pharmaceutical product to be licensed, it is understandable
(even if regrettable) that the manufacturers do not engage in this research.
In the Netherlands, steps have been taken to address this potential problem. In
1997, the Paediatric Pharmacology Network was founded to stimulate clinical
research on paediatric prescriptions, but this network has scant personnel and
financial resources. In 2001, the Health Inspectorate (Inspectie voor de
Gezondheidszorg) called for research, especially on psychoactive drugs, and
called for the establishment of a database of drug-related incidents in the
paediatric population. Similar calls have been issued by, among others, the
European Agency for the Evaluation of Medicinal Products, resulting in a European
Network for Drug Investigation in Children. The European Commission has also
shown awareness of the issue. However, results of this movement are slow in
coming.
The Dutch Health Care Insurance Board (College voor Zorgverzekeringen, or CVZ)
issued a call for tenders in late 2000, requesting a study to prioritise research
needs in the area of prescribing for children. RAND Europe responded with a
proposal to conduct a review of the scientific literature on this topic and to
disseminate a survey to family practitioners, paediatricians and pharmacists to
obtain their perceived needs for clinical research in this area. The study was to
be guided by a group of experts in the field of paediatric prescription,
representing the various medical specialities involved. This proposal was
accepted by CVZ, and the study commenced. However, as the project unfolded,
the guidance group expressed the opinion that the original research question was
incomplete, because the problems in prescribing for children could not be
resolved by clinical research alone; other issues were also important. In
x
response, RAND Europe proposed, and the CVZ accepted, a modified study to
identify the barriers to effective prescribing for children, with an emphasis upon
the absence of good clinical research within those barriers. In the remainder of
this summary, we will discuss the study as modified.
The goals of this study are threefold.
•
The first is to bring to awareness and to assess the importance of barriers to
optimal pharmacotherapy for children. Here, we take the term barriers in its
broadest sense, to include shortcomings in knowledge, communications,
individual caregiver and patient behaviour, and the organisational structure of
the health care delivery system.
•
The second goal is to identify possible solution directions—or
recommendations for change—that can reduce these barriers. These solution
directions are also assessed in terms of their likelihood of reducing barriers,
with the assessment based on the likelihood of the solution being successful
for the number and importance of the barriers to which they would be
addressed.
o
A subgoal of the second goal is to assess the extent to which a new
organisation focussing on prescribing for children, would be one of
the promising solution directions.
•
The final goal of the study reflects back to the original purpose of the project,
and is to assess the extent to which clinical research is a useful solution
direction, and which aspects of such clinical research (specific drugs, clinical
condition addressed, dosage, side effects, etc.) are most important.
Methods
The study began with an examination of the literature on paediatric prescriptions
and with a series of interviews with different stakeholders in the health care
delivery system. This resulted in a survey sent to family practitioners,
paediatricians and pharmacists, followed by a second round of (group) interviews
with (a different set of) stakeholders. These led in turn to a set of quantitativelyposed questionnaires sent to the guidance committee assessing the discovered
barriers, possible solution directions, and clinical research priorities. The results
of these questionnaires were presented and discussed in a meeting of the
guidance committee.
Orienting interviews. In parallel with the review of the literature, a series of
orienting interviews were held with organisations that have an important role in
xi
health care for children. The purposes of these interviews were to learn what the
roles of these organisations with respect to paediatric care were and to gain
stakeholder perspectives on the possibility of a new organisational structure.
Review of the literature. The review of the literature followed the wellestablished methods for systematic reviews. Search terms included medication,
p(a)ediatric, child, off-label, unlicensed, pharmacol*, clinical trial and review.
The search was constrained by examination of the prevalence of certain diseases
among children and the extent to which certain medications are prescribed for
children. To limit the scope of the study, articles regarding prescription for
neonates and children under the age of 3 months were excluded. On the basis of
these criteria, the following disease categories were examined:
•
Allergies and immunological disorders (especially asthma and allergic
rhinitis)
•
Infections (otitis media, respiratory disorders and urinary tract infections)
•
Endocrinological disorders (diabetes, growth hormone deficiency,
congenital adrenal hyperplasia, hypothyroidism)
•
Psychiatric disorders (ADHD, autism, depression, enuresis)
•
Neurologic disorders (epilepsy)
From an original set of 30000 "hits", a filtering process reduced the set of articles
examined to a more manageable but still sizeable 600.
Postal survey. On the basis of the literature review and the advice of some
members of the guidance committee, a survey was designed and distributed.
The survey selected 22 medications on the basis of one or more of the following
criteria: (1) they were in the top-10 of medications prescribed for children, (2)
they were typically prescribed off-label, or (3) the literature review identified
them as "hot" items. For each of the medications, the survey asked whether the
respondent prescribed it for children, whether they believed they were adequately
informed regarding dosage, side effects and interactions with other medications,
and for which age groups they would like to have more information. Respondents
were also given the opportunity to add up to five medications not on the list of 22
that they believed were problematic for children. Finally, respondents were asked
to prioritise the (22 to 27) medications in terms of their need for clinical research.
The postal survey was suitably tailored for family practitioners, paediatricians and
pharmacists, and distributed to a sample of 200, 450 and 200 of each group,
respectively. The response to this survey was disappointingly small, with a 27.5
xii
percent response rate for the paediatricians, but only 12.5 percent for the other
two groups (in spite of follow-up attempts for some of the addressees). Given
this response, statistical analyses were performed only for the paediatricians.
Second round interviews. To obtain further insight into research priorities and
to identify other forms of barriers, a second round of interviews were conducted,
with a different set of individuals than in the first set. Two types of interviews
were performed, medical-content interviews from a total of 13 medical specialists
and pharmacists and policy interviews with a total of 9 policymakers from the
ministry of health, the pharmacological industrial sector, and self-standing
organisations concerned with medications and paediatric care. These interviews
were used to identify the various barriers to optimal paediatric prescriptions and
to get ideas how to resolve the barriers. The starting points for the interviews
were the results of the orienting interviews, the literature review and the survey.
The interviews were analysed to identify the major barriers mentioned, which
were then categorised (see results subsection, below). On the basis of ideas
presented in the interviews and the literature, plus the knowledge of the research
team, a set of possible solutions to the barriers were assembled and also
categorised. The barriers and solutions were presented to the guidance
committee for discussion and refinement.
Quantitative questionnaire. The refined lists of barriers and solution directions
were put into a questionnaire given to the members of the guidance committee.
The committee was asked to rate the importance of the barrier as an impediment
to paediatric care, and then to rate how effective each proposed solution direction
would be in alleviating each barrier. The committee was also asked to rank each
of ten categories of medications on the improvement of prescribing for children of
further knowledge development and dissemination of information on safety,
effectiveness and long-term effects of each category. The results of the first two
parts of the questionnaire formed the basis of the prioritisation of bottlenecks and
solutions, and the results of the third part formed the basis of a priority listing for
clinical research within the overall prioritisation of solutions.
Results
Here, we first identify and prioritise the barriers to effective prescribing for
children, followed by the same treatment of the possible solution directions. We
considered possible ways to implement the favoured solution directions. Finally,
xiii
we examined the types of medications requiring research and the nature of the
problem that the research should address.
Barriers. Table S.1 presents the 14 barriers that were identified, along with the
average importance by the guidance committee of each barrier as an impediment
to effective paediatric prescribing. The barriers are rated on a 1-to-5 severity
scale, where 1 is low severity of the barrier and 5 is great severity of the barrier.
Table S.1 Barriers to Optimal Prescribing for Children
Barrier
Severity
A. Lack of evidence-based knowledge
A.1. Lack of basic general knowledge on how medications work in
3.9
children
A.2. Lack of knowledge regarding specific medications
4.1
B. Barriers to knowledge development regarding
pharmacological research using children
B.1. Methodological problems
3.2
B.2. Legal problems
3.6
B.3. Political problems
3.5
B.4. Ethical problems
3.5
B.5. Financing problems
4.6
B.6. Inadequate registration of experience from practice
3.9
C. Barriers to knowledge uptake
C.1. Lack of clarity in the basis of information in guidelines, formularia,
3.5
etc.
C.2. Education of care givers
3.7
C.3. Lack of dissemination of existing knowledge
4.0
C.4. Off-label prescribing
3.8
D. Barriers to implementation of knowledge
D.1. Relationships between pharmacists and physicians
3.1
D.2. Contextual barriers
3.0
All of the barriers had a mean rating of 3.00 (the middle of the scale) or above,
indicating that each of them was considered by the guidance committee to be
real. The highest score was obtained by B.5. Financing problems and is a
reflection of the fact that nobody—the pharmaceutical sector, the government,
research institutions—has this research as a high spending priority. Second and
xiv
third-ranked, with scores in the 4.0 or above range, were A.2 Lack of knowledge
regarding specific medications and C.3 Lack of dissemination of existing
knowledge. The importance of barrier A.2 justifies the original intent of this
study—there are medications that require attention. The importance of barrier
C.3 says that even when the research is performed, it is not always viewed as
adequately disseminated.
Solutions. The guidance committee was asked to assess the effectiveness of 11
different solution directions on a scale from –2 (would make matters much worse)
to +2 (would significantly improve matters) for each of the 14 barriers posed in
Table S.1. For each respondent, the overall rating of each solution direction was
taken by calculating the average cross-product over the 14 barriers of the
contribution of the solution direction times the respondent's own rating of the
importance of the barrier. Thus, for each respondent, each solution direction had
a rating between –2.00 and +2.00. These were averaged over respondents and
are presented in Table S.2.
Table S.2 Solution Directions for Reducing the Barriers to Good Paediatric
Prescribing
Weighted
Solution direction
score
Governmental stimulation of paediatric clinical research
+0.8
Harmonise and reduce restrictions on paediatric clinical research
+0.4
Register and analyse paediatric clinical experience with
+0.6
medications
Consolidate knowledge in databases, guidelines, etc.
+1.0
Accept off-label prescribing
-0.3
Extend the time duration of patents
+0.2
Require research with children before registration of medications
-0.4
Require research with children within a time period following
-0.4
registration for adult use
Establish a knowledge centre for paediatric pharmacotherapy
+1.2
Extend the training of paediatricians to include pharmacokinetics
+1.0
Establish a national formulary for paediatric prescription
+1.0
The results of this questionnaire are clear; there is a consensus for increasing the
availability of knowledge through a knowledge centre, more focussed education,
consolidation of knowledge, and establishing a national formulary. Giving up on
xv
the problem by accepting off-label prescribing is not favoured, nor is requiring
research with children as a part of the licensing requirements for adult use.
Solutions offered to stimulate the pharmaceutical industry were generally not
favoured.
For each of the four top solution directions, we considered an effective way to
implement it, taking into account the content of the solution direction, the roles
relevant actors might play, the financial requirements, and possible
disadvantages that might be incurred.
Medications. Based on the survey results and the second round of interviews,
ten categories of medications were presented to the guidance group for
assessment:
• Behaviour modifiers and psychotropics
• Painkillers
• New anti-epileptics
• Growth hormones
• Anti-allergenics and anti-asthmatics
• Prednisone
• Medications for sleep disturbances
• Medications for gastro-intestinal complaints
• Antibiotics
• Anaesthetics
From zero to four specific examples within each category plus a "general" rating
for the category were assessed, for a total of 27 medications. The assessment
was on a 1-to-5 scale in terms of the need for knowledge development and
knowledge dissemination for each of three aspects:
• Safety (including drug interactions and side effects)
• Effectiveness (including appropriate dosage and route of administration)
• Long-term effects
Thus, there were 162 total ratings done (27 medications, development vs.
knowledge, 3 aspects). The results are presented in Table 6 (Chapter 4) of the
main text.
Considering the average score across all medications, dissemination of knowledge
(average score 3.4) was viewed as more important than development of new
knowledge (average score 2.8). Long term effects (3.3) were more important
than pharmacokinetics (3.1) or safety (2.9). Among the medication categories,
xvi
the high scorers were behaviour modifiers (especially Ritalin) and anti-epileptics.
These two categories were viewed as short on knowledge of long-term effects,
but countered the general trend across all medications by requiring more
knowledge development than dissemination. These findings also held for the
specific medications within these two categories of Topiramat and Lamotrigine
(anti-epileptics) and the behaviour modifier Ritalin.
Conclusions
The findings presented immediately above largely speak for themselves; the
study has identified significant barriers to optimal prescribing for children and a
number of solution directions that have promise for lowering multiple barriers.
The original impetus for the study—prioritisation of paediatric pharmacological
research—has been shown to be justified, although such research must be
conducted alongside other important steps. Within the prioritisation of research,
the study of long-term effects of behaviour modifiers and anti-epileptics—
medications that children may take for extended periods of time—received the
highest call. For medications where the need for knowledge development is not
so compelling, more attention should be given to knowledge dissemination.
However, the study also clearly showed that a large number of problems in the
context of medical content form important barriers to knowledge development,
knowledge dissemination and general good paediatric practice. Therefore, the
solutions offered for both the medical and the policy problems deserve attention.
How to implement the recommendations that flow naturally from these findings is
a different matter. The role of the Netherlands in this matter is not
determinative, but can be important. A unified approach in the Netherlands,
coalescing the government, health care delivery sector, medical research sector
and pharmacological industrial sector, can improve the situation in the
Netherlands and serve as a blueprint for harmonisation within Europe. Financial,
ethical, legal, and organisational aspects of the problem are best addressed in
this integrated cadre, rather than each sector pursuing its own interests in
isolation.
In summary, we have verified that the problem of suboptimal prescribing of
medications for children is a real problem, that the causes of the problem are
complicated, and that solving the problem calls for a scientifically and politically
integrated approach. With such an approach, paediatric prescription can move
from an art to a science, and all—but especially the children—will benefit.
xvii